The goal of this study is to remediate word-finding problems in patients with Primary Progressive Aphasia (PPA) and to delay the further progression of word-finding impairment. The current approach is novel in that it contains a prophylaxis component in which we attempt to strengthen neural connections that remain functional, making them more resistant to degradation as the disease progresses. While the study is specific in its targeting of word-finding problems, a successful outcome would bode well for other studies aimed at prevention or reversal of declining cognitive functions in dementia. Patients with PPA whose naming of pictures has just begun to show signs of decline will receive practice with picture naming in two conditions: viewing the picture and repeating the name; and viewing the picture with its written name, plus reading and writing the name. Naming of pictures trained in each of these conditions will be compared, at three time intervals post-training, with naming of pictures tested before the study but never trained. A scene-description task containing trained items will be used to assess generalization. It is predicted that the pairing of the picture with its written name, combined with the motor task of writing the name, will result in a greater ability to name the picture at a later date than simple practice viewing the picture and repeating the name. Generalization to other exemplars and to other task situations is also predicted. In addition, we will image the brains of the patients to determine whether patterns of brain atrophy can be used to predict treatment response.